Evolving cultures: Shared intentionality and the evolution of symbolism

1 Introduction

One question philosophers have dealt with for ages concerns the definition or supposed “essence” of the human being. While of itself an old question, it is today often associated with the philosophical anthropology movement of the early twentieth century, led by such figures as Max Scheler (1874 – 1928), Ernst Cassirer (1874 – 1945), and Helmuth Plessner (1892 – 1985), who interacted in different degrees with phenomenologists like Edmund Husserl (1859 – 1938) and Martin Heidegger (1889 – 1976).

More recently, the question what defines the human being has been regaining some attention in debates in cognitive archaeology, albeit in modified form, and often in light of new discoveries pertaining to our Neanderthal cousins. Debates now usually focus on the application of such concepts as “behaviorally modern” and “modern symbolism”. These closely related concepts are taken to capture something distinctively human, with their application to earlier, usually Middle Stone Age/Middle Paleolithic (MSA/MP), hominins being the central point of controversy.

Archaeologists tend to associate concepts like “modern symbolism” with the production of certain kinds of seemingly non-functional objects or aspects of them. For most of the twentieth century, these and other, similar terms were standardly reserved for the sort of behavior associated with Late Stone Age/Upper Paleolithic (LSA/UP) figurines, such as the famous lion-man of Hohlenstein-Stadel, dated some 30kya (Dalton 2003), as well rock art, such as of Chauvet and Lascaux, dated some 28kya-37kya and 17kya respectively (Quiles et al. 2016; Ducasse and Langlais 2019). These were, of course, all made by Homo sapiens, and it is worth noting that to date, there is no evidence of any comparable depictive art by any other species of humans, or by MSA Homo sapiens in Africa.

Nevertheless, more recent findings have challenged the classic picture of a “creative explosion” (Pfeiffer 1982) in the LSA/UP. Use of pigment may go back 400 – 500kya (Wynn 2012, 290 – 291; Rifkin et al. 2015; Dapschauskas et al. 2022), and its decorative use is often taken to indicate modern symbolism as well, although pigment also has various functional uses (Henshilwood and Dubreuil 2009, 50; Vanhaeren and d’Errico 2006, 1107). The role of beadwork, a more recent invention but nonetheless originating well before the LSA/UP and not unique to Homo sapiens, is now also often said to be a symbolic one (Bouzouggar et al. 2007; Zilhão 2007; Zilhão et al. 2010, 1023; Henshilwood and Dubreuil 2011, 375; d’Errico and Vanhaeren 2012; Prévost et al. 2021, 1). In addition, the possible decorative use of pendants and eagle talons, also by Neanderthals, might suggest behavioral modernity (Rodríguez-Hidalgo et al. 2019; Frayer et al. 2020), and recent research suggested the decorative preservation of lion claws by Middle Paleolithic hominins as far back as 190kya (Russo et al. 2023).

In light of such findings, some authors now suggest that modern symbolism is not underpinned by a distinctive genetic and/or cognitive basis unique to Homo sapiens (Wolpoff et al. 2004; Zilhão et al. 2010; Colagè and d’Errico 2018). Recent evidence pertaining to ochre, pendants, and beadwork from Neanderthals has been taken to indicate that behavioral modernity “is not a species-specific phenomenon” (Zilhão 2007, 1), and that Neanderthals and modern humans must have been cognitively indistinguishable. Others have been more critical of such assessments, however, and the view remains represented by a minority today, according to Papagianni and Morse (2022). Indeed, while there is increasing evidence for comparably complex tool production of Neanderthals and Homo Sapiens for much of the earlier MSA/MP, evidence for the use of beadwork and pendants by Neanderthals is still relatively sparse, and the absence of Neanderthal depictive art should not be ignored. Also, some research suggests Neanderthals may have had better auditory and speech capacities than expected, but still not quite in the modern human range (Conde-Valverde 2021), and other recent work points to differences in neurogenesis (Pinson et al. 2022).

Everyone, then, agrees that the cognitive and behavioral gap between various MSA/MP hominins is smaller than previously thought, but there is little more agreement than that. This lack of consensus is not due only to a gaps in data. Debates are also said to suffer from “loose reference to symbolic culture” (Wynn et al. 2016, 15) and a “poverty of appropriate interpretive concepts” (Wynn and Coolidge 2010, 5). Indeed, there is no consensus as to what “modern symbolism” consists in, and there is a tendency to reduce the complexity of archaic hominin behavior to simple dichotomies, such as between “modern” versus “pre-modern” behavior, which reflects the essence-thinking of the old philosophical anthropology.

A closely related disagreement in the literature is the one between advocates of continuity and discontinuity approaches. Continuity approaches emphasize that complex human behavior evolved gradually, often involving a mosaic of different elements which gradually came together. Thus symbolic capacities may have become “gradually richer and more elaborate […] without any sudden changes or truly novel features” (Parravicini and Pievani 2018). From this viewpoint, there need be no “essence” of the human being which sets it apart from other, extinct hominins. Discontinuity approaches, on the other hand, tend to emphasize the importance of a single, usually genetic, change – what Bickerton (2000) called “factor X”. Discontinuity approaches tend to fit better essentialist thinking about human nature, as they suggest archaic hominins missed a certain “factor X”, which determined the success of Homo sapiens.

In the following section, I first reflect on some of the conceptual problems involved in the debate on modern symbolism, which in my view cast doubt on the usefulness of generic concepts like “modern symbolism”. In the following three sections, I take a step back, and provide a basic outline of the mechanisms of the evolution of material culture. Here I argue that macro-cultural change can be viewed in terms of the selection of useful adaptations which are transmitted across generations, as in natural evolution. There are differences between both evolutionary processes, however. In particular, I suggest that culture relies mainly on so-called shared intentionality, which makes horizontal and vertical transmission possible through what I conceptualize as generative pathways. This transmission process does not rely on genes, but on the brain’s plasticity and to a lesser extent external vehicles. With this fairly basic account of cultural evolution at hand, I return to the debates about (dis)continuity and modern symbolism in sections 6 and 7. Here I defend stage-based approaches to the evolution of material culture, as they offer gradualist alternatives to the dichotomous thinking dominating current debates, and better fit the theory of cultural evolution outlined.

2 Modern symbolism: Basic problems

In this section, I consider some basic problems concerning the concept of “modern symbolism”. To start off, each of these two words currently lacks an agreed upon meaning. The concept of “symbolism” holds connotations to that of a symbol, which taken in the familiar Peircean sense, means a signaling relation which is both arbitrary and conventional (see van Mazijk 2023). While this definition of a symbol applies, for example, to the representations involved in writing, the use of ochre, pendants, beadwork, or even figurines and rock art do not entail symbolism in this sense. Decorative ochre use, pendants, and beadwork may be said to signal something, but it is far from obvious that they hold conventionally determined reference. Figurines and rock art, on the other hand, are icons. It is thus from the start unclear what the concept of “symbolism” is doing in discussions on ESA/LP and MSA/MP material culture, for none of the items of these periods clearly carry information in the way symbols do. Until a better concept is found, a simple improvement would be to speak of material culture instead.

It is also unclear what the “modern” part is supposed to refer to, and authors often do not specify whether the material culture itself, or the implicated behavior, genetics, or cognition^[1] is supposed to be at stake. As also noted in the introduction, the concept of “modern” may (inadvertently) evoke an image of a dividing line between modern and pre-modern behavior, which in turn may motivate unproductive semantic discussions.

From the perspective of cognitive archaeology and/or evolutionary psychology, we not only want to understand material culture, but also the cognitive or behavioral changes it implies. In this regard, it is useful, first of all, to distinguish between the roles of biology and culture (without implying they are radically distinct). To give an example, MSA/MP beadwork and pendant construction may involve a wide range of genetically fixed cognitive abilities, like memory, imagination, navigation, sensorimotor skills, and so on. Cognitivist and neuroscientific approaches in archaeology may help elucidate the role of such abilities. However, besides this, production most likely involves a certain inauguration into a culture. I define culture more precisely in the next section; for now, it can be taken quite generally to concern the sort of learning and communicative practices by which humans share ideas, skills, knowledge, plans, etc.

One important feature of culture is that it can be passed on to the next generation. That is to say, culture allows of vertical transmission, generating a link between living generations and their ancestors. As (archaic) humans transmit ideas from one generation to the next, the cultural achievements of past communities remain effective for them, even those of many generations ago, although often in modified forms. Such a link, which connects humans across generations, can be conceptualized as a generative pathway. Generative pathways need to be sustained by continuously passing on cultural achievements from one generation to the next; any break would mean an abrupt loss of culture. However, if appropriately maintained, culture might also accumulate, as new generations can build on the achievements of previous ones.

The importance of generative pathways is sometimes overlooked, but it can hardly be overstated, as humans are – as Husserl also suggested in the opening quote of this contribution – virtually nothing outside of the unbroken chain which connects them to their ancestors (Henrich 2017). For example, although largely based on anecdotal evidence, modern humans growing up outside of society and the continuous chain of generativity which connects them to previous societies (so-called feral children), appear so severely impaired that they cannot even develop language (see discussion in Candland 1993, 9 – 26). Likewise, it is virtually inconceivable that a contemporary human, let alone an MSA/MP hominin, could spontaneously invent beadwork production, that is, without any access to generative pathways. This art presupposes, then, besides genetically fixed brain functions, the development and maintenance of generative pathways, through which the vast array of accumulating techniques and cultural behaviors required to produce such tools is transmitted.

Human behavior depends thoroughly on cultural transmission, and this dramatically complicates our understanding of prehistoric hominin behavior. For one, it makes it very difficult to determine how hominins produced their material cultures. One might think that we should be able to identify certain necessary and sufficient conditions for the production of tools like beadwork. But this need not be the case. For example, it is perfectly conceivable that some MSA/MP hominins relied on improved processual memory, or any other genetically fixed cognitive function, whereas other hominins relied more substantially on culture and instruction. Both groups, then, would produce the same tools, but the production would be realized in different ways. This problem, sometimes called equifinality, suggests that complex or modern behavior need not be grounded in any particular set of necessary and sufficient conditions.

A related complication is that the same material products may also realize different functions, and therefore could point to different behaviors. While on a holiday in Athens, I asked a friend why she wore jewelry. She told me that she wore one piece because her grandfather gave it to her; it made her feel connected to him. Other jewelry she wore only with certain clothing on befitting social occasions, and other jewelry all the time because it looked good. This illustrates nicely that cultural objects are not only possibly multiply realizable, but they might also themselves realize multiple functions. Something similar may apply to MP/MSA beadwork and pendants, and this again complicates our understanding of use and function. Thus, the multiple realizability of products (in terms of their biological and generative conditions) and the multiple realizability of the functions of those products (in terms of their socio-cultural use) pose significant interpretative problems for archaeologists.

Other problems concern the descriptions we use to assess use and function. First, it is not always clear what sort of understanding otherwise intuitive descriptions yield exactly, and we sometimes shift without noticing it between explanatory discourses. For example, from the perspective of social cognition, beadwork might be said to aid social cohesion, from a biological perspective to enhance chances of survival or reproduction, and from a phenomenological perspective the ability to express oneself. Each of these suggestions may be informative regarding the cultural evolution of beadwork, but these descriptions also clearly overlap. Consider the remark by Donald (1991, 10), that “it could well be the case that the intellectual capacities needed to sustain large groups are identical to those that enable cultural invention”. In saying this, he probably had in mind that the same “intellectual capacities” could have different functions. However, it could also be that “sustaining large groups” and “cultural invention” are not in fact two radically different things to start with, but to an extent overlapping descriptions.

Humans, then, may not always be well-equipped to think about complex behavior, and our concepts and explanations are often quite vague and overlapping. This holds in particular for broad concepts, which are inherently less likely to represent any particular function. This may well hold for the concept of language, which is traditionally conceptualized as a single “symbolic function” (Cassirer 1972), and it almost certainly holds for concepts like “art” and “symbolism”.

Finally, our assessment of use and function is complicated by the many difficulties involved in considering changes over long periods of time. To give just one example, there is the possibility of exaptation, that is to say, of changing use and function. This can hold for physical traits, but also for material culture. For example, it is possible that the cultural use of ochre was first borne out of practical needs, and gradually evolved to carry social information. The adaptive function of ochre use then changed over time, a difference which need not be visible in the archaeological records.

These were just a few of the problems concerning the use of concepts like “modern symbolism”, and I skipped many others, as those pertaining to the correct use of inference, conflicting data, and cultural biases. Many of these problems apply to all prehistoric interpretation, and they need not stop us from making sense of the deep past. However, it seems to me that a concept like “modern symbolism” faces more substantial problems, as this concept is currently too broad and vague. Focusing discussions on modern symbolism (or modern behavior for that matter) may do more harm than good, as it distracts from a more careful interpretation of data. One way to remedy this situation is to speak of “material culture” instead, and to try to make sense of the many observable changes in material culture over the long course of the Paleolithic – rather than to ask when behavior first became “modern”.

Thinking about cultural change, rather than focusing on the question of “modern” behavior, presupposes a viable concept of culture as well as of cultural change. In the next section, I outline a framework for thinking about culture, and section 4 and 5 offer a theory of how cultures might evolve. After these discussions, I return to the topic of modern symbolism and the debate about (dis)continuity.

3 An account of culture: Shared intentionality and generative pathways

In the previous section, I considered some of the problems involved in debates on modern symbolism. Along the way, I introduced the concept of generative pathways, which refers to the transmission of ideas, techniques, and behaviors – in short, of culture – across generations. It seems prima facie likely that an understanding of the mechanisms underlying this process of transmission, and therefore of cultural change, is important in interpreting observable changes in material culture in a scientifically responsible manner. Among others, an understanding of macro-cultural change could inform debates on (dis)continuity and (pre)modern symbolism. Yet very few cognitive archaeologists seem to engage with this topic. Before developing an understanding of macro-cultural change in section 4 and 5, I first defend a particular take on culture in this section.

There is a relatively strong “Platonic” tradition in philosophy living to this day, which views the cultural achievements of humankind as “spiritual”, in contrast to “natural” (Geist and Natur, as the German called it, e. g. Husserl, Scheler, Plessner). This realm of “spirit” has numerous analogs in Anglophone philosophy today, for example in the so-called “space of reasons’’ of John McDowell (McDowell 1994; 2009; 2013; van Mazijk 2020). Many of such contemporary analogs do not completely separate spirit from nature, but they do at least maintain this distinction, as a way to separate humans from non-human animals, or to conceptualize the subject matter of the humanities in contrast to that of the natural sciences.

In my view, the concept of culture can be significantly de-Platonized (naturalized) by defining it as the product of a certain type of social cognition, one that is characterized by what is today called “shared intentionality” (Tomasello 2010; 2021; van Mazijk 2024a; 2024b; 2024d; 2024 g). Shared intentionality involves the understanding of being directed at a third thing together, rather than merely individually. For example, humans and some captive primates like chimpanzees can manipulate each other’s (or a caretaker’s) attention by pointing. Generally speaking, pointing has a referential and a social intention, the latter usually being some request. Interpreting this social intention (“what does the other want me to do?”) requires more than merely tracking the direction of pointing. In particular, it involves a complex recursion of intentional states, today commonly called third order intentionality, as in “I see that you want to show me X” (see also van Mazijk 2024a; 2024b; 2024d; 2024 g). This contrasts with, for instance, a more basic act like gaze following, which involves two orders of intentionality (“I see that you see X”).

Put in simpler terms, sharing intentionality requires what the philosopher Edmund Husserl called “reciprocal empathy” (Husserl 1973b, 471 – 479). That is to say, it includes a complex awareness of each other’s intentional states, in such a way that there is awareness that both subjects are directed at some third thing, hence together, rather than merely individually and contingently. According to Husserl, this includes two subjects, S¹ and S², which are together directed at a third thing: “S¹ grasps empathically S², and S² grasps empathically S¹, but not only that: S¹ experiences (understands) S² as experiencing S¹ understandingly, and vice versa” (Husserl 1973a, 211). This reciprocal empathy thus involves what is today called third order intentionality (Tomasello 2010; 2021), where both subjects are not only directed at X, but moreover entertain something like “I see that you see that I see X”.

This complex recursion of intentional states allows minds to connect and to share object-directedness, which opens the door, or so I argue, to a new kind of information exchange. Any information thus exchanged might be inaccessible to any individual – including non-human animals and infants – who has not at some point partaken in the relevant processes of sharing attention. For example, a bison or a human newborn do not understand words, they do not see the practical function of a Levallois stone tool, and they do not grasp the meaning of the symbolic signs engraved in a Mesopotamian clay tablet, for the simple fact that neither has partaken (and the bison can never partake) in the right social intentionality. Words, tools, and symbolic signs, then, presuppose a capacity for mutual empathy and shared object-directedness. Their “meaning” is group-bound, and only appears to those who have been appropriately socialized by sharing attention.

It makes sense, in my view, to define “culture” primarily in terms of the totality of group-bound behaviors and significances which are originally^[2] accessible only to those who have partaken in the relevant processes of shared intentionality. Culture then depends, in one way or another, on the psychological principles of sharing intentionality. I realize that this is a somewhat demanding concept of culture, as it excludes the products of mere imitation, which do not presuppose shared intentionality, as one individual can imitate another who is not socially engaging.^[3] For example, extant non-human primate tool use (like nut cracking) appears based on imitation, not instruction, which would require sharing intentionality.

It is true that imitation also requires complex social understanding. In particular, it requires second order intentionality, of the sort “I see that you see X”. Imitation does not, however, presuppose the mutual recognition of being similarly directed at a third thing, which is what shared intentionality is about. Non-human primates, at least in the wild, do not actively form such triadic bonds. They do not point out things to each other, so as to be directed at something together, as a “we”. They also do not actively give names to things by pointing them out to each other, to create new meaning structures which are group-bound, and they further do not (clearly) actively instruct each other in practical activities. Instead, they mainly communicate through pantomime, which is dyadic (see e. g. Pika and Mitani 2009; Gibson 2012; Tomasello and Call 2019; Tomasello 2021). Interestingly, chimpanzees are in fact capable of pointing, which serves to establish shared intentionality, in captivity. Consequently, they appear to possess the relevant cognitive infrastructure, while lacking the right (cooperative) motives to use it in the wild, as Tomasello (2021) has also argued.

Clearly, some form of information transmission is possible based on second order intentionality, as in imitation. However, a concept of culture based only on shared intentionality does a much better job at capturing what’s distinctive about human cultures, as opposed to the “cultures” of other animals, which tend to rely on imitation (e. g. nut cracking). Only humans actively engage in shared intentionality, and only humans thoroughly depend on cultural transmission for existence – although this skill and human dependence on it may certainly go back millions of years (see Henrich 2017; van Mazijk 2024d).

The concept of shared intentionality, then, while being somewhat restrictive in excluding most animal “cultures” based on imitation, may go a long way in addressing the gap which exists between non-human animal cultures and human cultures. An additional advantage of basing culture on shared intentionality is that this type of social understanding does not invoke a hard dividing line between humans and other animals. For, as mentioned, even chimpanzees can in fact engage in shared intentionality in captivity, when motivated by sufficiently cooperative human caretakers (Leavens et al. 2005). Viewing culture in terms of shared intentionality thus conceivably bridges the gap between human and non-human primate cognition, making it easier to see how culture evolved gradually out of ape-like intentionality, with its likely starting point 2 – 3mya ago (van Mazijk 2024d). By contrast, the more commonplace (at least in archaeology) conception of culture in terms of material-symbolic culture risks inadvertently emphasizing the gap, since detailed material cultures appear much later in the records, with no clear analog in non-human primate “cultures”.

My suggestion so far, then, is that we view material culture (like tools) as well as abstract culture (like verbal or gestural communicative signs), as depending largely on a specific kind of social understanding, namely the one involved in sharing intentionality. One might also simply say that culture requires a certain type of social bond, namely we-groups, based on sharing attention, which differ from the dyadic (I-you-groups) characteristic of other primate social life. The formation of such we-groups allows individuals firstly to transmit culture horizontally, that is, between co-existing individuals. However, we-groups can be composed of constantly changing individuals, and this effectively allows humans to pass on culture vertically as well, to subsequent generations, even when there is no intention to do so. For example, one individual might teach others in its group a new way of producing tools by sharing intentionality (showing objects, pointing them out to others, fixing gazes by speech acts or attention-getters, etc.). This particular know-how may subsequently outlive the inventor and their pupils, as long as it is consistently carried over onto new subjects who continuously join the group. This way, sharing intentionality becomes formative of what I earlier called a generative pathway, which links individuals across generations, in terms of cultural accomplishments.

It appears that, probably quite early on in our evolution, the advantages of shared intentionality and the culture it yields made hominins thoroughly dependent on these generative pathways. Put differently, culture, as the product of sharing intentionality, became an integral part of the ecological niche hominins inhabited, and to which new generations had to adapt in order to survive (Henrich 2017). For example, it has been suggested that ESA/LP hominins’ knapping skills indicate “evolving social cognitive capabilities” (Stout and Semaw 2006, 317) as well as teaching and practice (Hiscock 2014, 27). Morgan et al. (2015) further hypothesize that these Oldowan hominins subsequently became dependent on this very culture-infused ecological niche, such that “reliance on Oldowan tools would have generated selection favouring teaching”. In short, generatively transmitted, cultural behavior (Oldowan tool-making) changed the ecological niche hominins inhabited, to which they subsequently adapted biologically speaking.

Morgan et al. (2015) seem to have had mainly biological adaptations in mind. In this sense, culture becomes another environmental factor which exerts selection pressure on heritable traits, and thereby becomes part of the story of evolution by natural selection. As Blackmore (2010, 268) also notes, “brains gradually became better and better at copying [with culture]”, and Ermakov and Ermakov (2021, 3) suggest in a similar vein that “changes in cultural inheritance can lead to changes in genes”.

However, it is sometimes overlooked that hominins may also have adapted to their culture-infused ecological niches through more culture. This is a quite different kind of adaptation, which is today still poorly understood. For example, we might be able to consider a useful idea, a new word, a bond-strengthening item, or a new tool-making technique as itself an item selected for heritability, not within genes, but within generative pathways, for the adaptive advantages they may yield. In other words, any product of culture, if sufficiently useful within an already culture-infused environment, might be viewed as being “selected” and “transmitted” to the next generation. This could then happen quite independently of any natural heritability (genes), for the transmission process is a completely new one, as it relies on shared intentionality.

In summary: culture need not be understood as something spiritual which floats freely above nature, but we also need not explain it away. Instead, culture can be viewed naturalistically in terms of social intentionality theory, that is: the complex understanding minds have of other minds, and more specifically, the mutual recognition of being so-and-so directed. This type of social understanding allows the transmission of ideas, techniques, and much more across generations. This brings us to the topic of macro-cultural change which I discuss in the next two sections.

4 Evolving cultures: Sharing attention and plastic brains

Evolution is often viewed as the change in heritable characteristics of a biological population (a species) over successive generations. Any particular species inhabits an environment (an ecological niche), and as this environment continuously changes, the species adapts under selection pressure. The evolution of any species, then, depends on the niche-relative selection of favorable heritable traits. This means that there is no more or less evolved species in an absolute sense; a species merely adapts to the ecological niche it inhabits.

While this account of evolution offers a good starting point, it excludes macro-cultural change – at least insofar as “heritable traits” was earlier taken in the usual, biological (genes) sense. For large-scale developments in culture, such as those in our recent (pre)history, do not seem to presuppose genetic change. Human cultures have evolved at an incredible pace for at least the past 40ky, while relatively few relevant genetic changes appear to have been centrally involved in this process. This raises the question: how does cultural change compare to the familiar Darwinian theory of evolution by natural selection?

There is today lively debate about cultural evolution, and about the extent to which principles of genetic evolution apply to cultural evolution (Boon et al. 2021). In my view, macro-cultural change can be regarded as a different form of evolution, which is both similar to and different from natural evolution (note that while I oppose “cultural” to “natural” evolution, I am not implying the former is not natural). This process can be seen as centrally involving at least the following three factors (see also Mesoudi 2011, 25 – 54 for a more elaborate introduction to cultural evolution, but without the concept of shared intentionality).

First, cultural transmission relies on sharing intentionality. Put simply, in contrast to genes, products of culture^[4] are not transmitted (as for most animals at least) through sexual reproduction, but by sharing attention. This makes the interactions between the individuals involved a crucial factor in cultural evolution. Culture is the product of different individuals who in mutual social understanding cooperate, instruct each other, and organize themselves socially. This means demographics and population density are important factors in cultural evolution. In some cases, behavioral differences reflected in the archaeological records may well be explained by differences in populations and demographics, or better, by differences in the generative pathways which these groups maintain. In particular, the possibility of interactions between groups can lead to a fusing of generative pathways, which can speed up cultural evolution. This may have been a crucial factor especially in early African Homo sapiens during the MSA/MP, and likewise in the cultural explosion in Upper Paleolithic Europe.

Second, the primary vehicles underlying these social activities are not genes, but brains. The internal storage of culture in the brain is sometimes referred to as the memotype (Heylighen and Chielens 2009). Cultural evolution depends largely on the brain’s plasticity: it’s capacity to be programmed with different “software”, which is made available by we-groups and their generative pathways. As Donald (1991, 14) once put it, cultures “restructure the mind […] in terms of its fundamental neurological organization”. Culture, then, presupposes highly plastic brains, which can adopt various cultural “software”. Lower degrees of neural plasticity and shorter infancy phases (themselves matters of genetic “hardware”) will limit the cultural impact of sharing intentionality, as would presumably have been the case with many archaic hominins.

Third, many, although not nearly all, cultural products involve materiality. The external storage of culture is sometimes referred to as mediotype (Heylighen and Chielens 2009). For instance, material tools are inherently made from external materials, and some information about tool-making techniques can be stored in these tools themselves. This holds even more for written language, which is itself little more than an external storage system. Culture, then, is not stored only in brains, but also in the material environment.

At the same time, not all products of culture can be said to be material or “extended” in this way, and this puts, in my view, limitations on theories focusing predominantly on material-external vehicles, like material engagement theory (Malafouris 2013). For example, symbolic speech acts also presuppose shared intentionality (shared directedness at abstract objects or objects which are named), and words must be transmitted through generative pathways. Yet words, and symbolic gestures alike, do not require external vehicles in the way writing does. It is in my view quite likely that rich material-symbolic cultures are relatively late cultural adaptations, while cultural evolution itself started much earlier (I am thinking well over 2mya, but there is no hard evidence for this; see van Mazijk 2024d where I elaborate this claim further).

In conclusion: there are, on the simplified picture presented, two different forms of evolution. I skipped many technical differences and ongoing debates about them, but the main differences between the two forms of evolution concern, in my view, the vehicles of the entities which are copied (brains versus genes) and the transmission process (social intentionality, in particular shared intentionality, versus sexual reproduction). Simply by sharing attention, (archaic) humans can transmit vast stores of cultural information across generations. In doing so, each generation reprograms the brains of the next generation. Given the right circumstances, this novel copying process can lead to significant change much faster than evolution by natural selection can account for.

Granted that this is all somewhat acceptable, at least as a very rough sketch, we have yet to address whether cultures also literally evolve, namely through mechanisms of selection and adaptation, as is the case in evolution by natural selection. In the next section, I try to further clarify the workings of macro-cultural change, and I suggest that this indeed involves mechanisms of selection and adaptation. After this discussion, I return to the topic of (dis)continuity and modern symbolism, to see how these reflections may inform our thinking there.

5 Evolving cultures: Cultural adaptation and selection

As the previous section outlined, culture is “heritable”, albeit in a completely new sense, namely through the principles of shared intentionality and the generative pathways they produce. Simply by sharing attention, (archaic) humans can transmit cultural information across generations. Importantly, in so doing, they not only constantly reprogram brains, but also change their cultural and material environments. For example, a world with new, complex tools in it creates a different environment or ecological niche, which will in turn favor some new ideas more than others. Perhaps, then, cultural change can itself be regarded as subject to niche-relative selection, in such a way that we should speak not only of gene-culture coevolution, but also of culture adapting to culture.

While cultural evolution theories are gaining grounds, they are still controversial, as they contradict longstanding ideas about the nature of art and culture. In particular, culture and art are often seen (i) to have non-functional aspects, and to (ii) result from subjective (free) creative processes. In what follows, I defend the idea of macro-cultural change by niche-relative selection against some possible objections, in a non-technical manner (see also Heylighen and Chielens 2009 for other objections, as well as replies to them).

First, it is sometimes supposed that the observable variations between human cultures is indicative of some sort of freedom, and that this variation must therefore be incompatible with cultural evolution by niche-relative selection. This is not true, however, for environmental pressure always allows variations to exist, and many differences between cultures are simply such variations. For example, the fact that humans all look somewhat differently is due to the fact that facial features are apparently not subject to a sufficiently strong selection pressure that could eradicate such differences. Something similar holds for many cultural differences. It should in this regard not be overlooked that human cultures are in fact often quite alike, just like human faces are still broadly similar. For example, eating rice with sticks or a fork is a cultural variation between populations which selection pressures allow to exist. But there is still a general cultural adaptation underlying both behaviors, namely the use of simple tools for food consumption. This cultural adaptation, then, was sufficiently beneficial to be selected in most populations (although not everywhere), and was transmitted through shared intentionality.

A second possible objection is that some cultural behaviors do not seem to advance survival or reproductive chances; they rather seem to be non-functional, or even to work contrary to reproduction. Such behaviors, for example institutionalized celibacy, are not, then, like forks and chopsticks, which are evidently useful. Such seemingly non-adaptive behaviors, like those of celibate monks, can usually be explained quite simply by considering their adaptive function within the larger population.^[5] While celibacy may not enhance the reproductive advantages of certain individuals, it can – and evidently has been – selected in different populations as a useful adaptation, for it is only a small part of a larger group which remains reproductive.

A third possible objection concerns the cultural impact of the actions of individuals. Let us take an influential individual like Vladimir Putin as an example. It appears to us that Vladimir Putin is a free individual, yet his decisions may have substantial effects on human cultures across the world. Such effects may well seem quite sudden, rather than gradual. Also, they could be interpreted by some as destructive, which does not go well with the idea that they would have been selected for being useful adaptations. How, then, do we make this compatible with niche-relative selection? Do the ways in which cultures change around us not reveal that culture is not a matter of selection and adaptation after all?

It is quite easy to be misled by the details of our everyday lives when considering principles that govern change over large populations and time spans (I am only defending macro-cultural evolution). The actions of any given animal are never fully determined by principles of selection, and Darwin himself certainly never suggested such a thing. Just as any individual bison may run itself off a cliff, Putin’s individual actions are just that: an individual’s actions, although here they happen to impact the lives of billions of other people as well.

Not all cultural change, then, is necessarily fixed by niche-relative selection. In fact, we do not ordinarily use the term “culture” to refer to useful adaptations at all. In everyday parlor, the word “culture” rather refers precisely to cultural variation: the different cooking styles, clothing, decorations, and so on of various people. Phenomenologically speaking, that is, from the human viewpoint, cultural variation seems to be far more salient than the adaptive patterns underlying them.

While such variations certainly remain of interest from the viewpoint of “micro-level disciplines” (Mesoudi 2011, 52), as in the arts and humanities, the study of hominin evolution plays out on an entirely different scale. Here, material (cultural and symbolic) changes are better understood in terms of functional adaptations, by considering how earlier humans adapted and evolved under changing environmental pressures. The environments partially change because of new cultural inventions, and new cultural inventions also result from changing environments.

In conclusion: by sharing intentionality, (archaic) humans transmit culture across generations, thereby changing their environments, which in turn changes their behaviors, which in turn changes their environments. This process of change is subject to niche-relative selection, and at least on a macro-scale, changes in material culture can be viewed as functional adaptations to changing culture-infused environments.

6 Continuity and discontinuity approaches

The previous sections provided a basic framework for thinking about cultural evolution. This framework bears on questions concerning the origins of symbolic modernity. I gave a fairly general account of what culture might be and how it evolves, which can occur both in “dialogue” with (gene-culture coevolution), but also independently of, genetic changes. In short, humans continuously adapt to their culture-infused environments, partially through culture, and they transmit such cultural adaptations horizontally and vertically by sharing intentionality. This process requires the brain’s plasticity, and may involve external vehicles. In recent (pre)history, with unprecedented population densities and complex fusions of generative pathways, this has given rise to incredible cultural variation, which is remarkably salient to the human eye. Over the longer course of hominin prehistory, however, culture need not have been so richly varied, and variation need not have been as salient to earlier hominins. In considering the longer process of the evolution of the human mind, cultural and biological adaptation (function), not variation, should be the principle concern.

This discussion on cultural evolution bears on a significant divide in discussions on the advent of “modern” behavior, namely between so-called continuity and discontinuity approaches. Some of the better known discontinuity approaches of the past decades have argued that single genetic mutations might be responsible for modern symbolism. For example, Chomsky famously maintains that there was a unique species-defining mutation, present in a metaphorical Adam or Eve, which resulted in behavioral and symbolic modernity (here guised as “infinite generativity”, Hauser et al. 2002, 1572). This mutation is taken to have occurred only in the Homo sapiens line, and no longer than 100kya.

In the field of cognitive archaeology, the enhanced working memory hypothesis by Wynn and Coolidge is quite well-known, and might be read as belonging to the same camp of discontinuity approaches. Wynn and Coolidge (2011) suggest that language alone is “insufficient to account for all of the features of the modern mind”, and that “several other components [were involved], including problem solving and long range planning abilities”. They subsequently hypothesize a possible mutation in the genetic or epigenetic basis of enhanced working memory, which could have been responsible for crucial changes in these cognitive capabilities (Coolidge and Wynn 2005; Wynn and Coolidge 2010). They further suggest that humans may have been neurologically modern sometime before this is reflected in the archaeological records.

A single mutation with a great impact on life on Earth is always a theoretical possibility, and it goes without saying that some genetic adaptations underlie species-specific behavior, including “modern” syntactical capacities (see Mozzi et al. 2016). Nevertheless, there are reasons, I believe, to be wary of interpreting “modern” human behavior exclusively along the lines of a single genetic mutation. This is firstly because, in the case of humans, a genetic explanation is never a complete explanation, because culture is no less a part of the story of our evolution, and it is, as I have argued, not reducible to genes. This means genetic explanation is bound to be one-sided. Second, it is worth noting that, during the whole history of life on Earth, there are only a few possible candidates for single mutations which completely changed the course of life on our planet, such as the first self-replicating RNA, the first mitochondria, and the first meiosis. Such events, then, are quite special, and only the most exceptional of circumstances should bring us to consider either natural or cultural evolution as discontinuous in this sense.

It is perhaps for similar reasons that Merlin Donald in Origins of the Modern Mind suggested that a continuity approach is “the only avenue open to us” (Donald 1991, 24). At the same time, he noted that “the evolution of cognition at the modular level might have continued well beyond the point at which physical evolution had stopped” (Donald 1991, 14). In other words, cultural evolution, from some point onward, may have started to completely outpace natural evolution, with human brains continuously being rewired independently of any genetic adaptations. Does this incredible acceleration of cultural evolution not suggest some kind of disruptive event after all?

In my view, we need not draw that conclusion. In this regard, it is sometimes overlooked that rapid cultural innovation is not in fact a universal phenomenon among humans from the LSA/UP onwards. Some smaller and isolated populations of “modern” humans, while often making more complex art and advanced tools than for instance the Neanderthals, did not accumulate culture as rapidly as in other, more densely populated parts of the world. From an archaeological perspective, some of them need not even appear as “advanced” as Magdalenian populations in Europe some 17kya. While these more isolated cultures could still be distinguished as “modern”, the suggestion of radical discontinuity appears motivated mainly by comparisons to humans standing in specific generative pathways.

At the same time, there is now increasing evidence for significant cultural adaptations in the MSA/MP. This is reflected, for instance, in research on MSA/MP symbolic speech capacities, as well as on the early use of decorative items and wooden tools. It thus appears that at least some archaic humans already developed the right genetic makeup as well as generative pathways required for relatively complex cultural practices. Such findings, too, blur the dividing line between modern and pre-modern behavior, and point to continuity.

Nevertheless, the old doctrine of a “cultural explosion” (Pfeiffer 1982) continues to exert a notable influence today, and this is partially due to the fact that there still is notoriously little evidence for significant material-cultural evolution during most of the ESA/LP, with only limited and often controverted evidence in the MSA/MP (such as decorative ochre use, the Excalibur axe, Shanidar “burials”, etc.). The archaeological records thus continue to sketch a picture of thousands of generations of pre-historic humans living in largely the same cultural worlds. In short, the archaeological records can still legitimately be interpreted along the lines of discontinuity.

However, it is worth repeating my earlier claim, that external materials are not the only – or even the main – vehicles for cultural evolution. Archaeologists have a long and detailed history of underestimating the graduality of cultural change, as they tend to focus one-sidedly on material culture. To give one arbitrary example, Steven Mithen, in his otherwise brilliant book The Prehistory of the Mind (1996), built a theory of cognitive modularity almost entirely around the gaps in the archaeological records at the time. He claimed, rather absurdly from a viewpoint a mere 20 years later, that archaic humans had a special cognitive module for stone manipulation, and could not see that other materials could also be worked. Likewise, Donald (1991), who based his ideas partially on the archaeological and paleontological data then available, hypothesized a (too) recent “mimetic phase” to fit the lack of evidence in his day regarding the speech and auditory capacities of MSA/MP hominins.

Obviously, theories should be based on evidence. Still, constraints of archaeological evidence should not, or at least not all too easily, motivate the construction of theories which are quite implausible from an evolutionary standpoint. In other words, archaeologists should take cultural evolution and its inherently gradual nature more seriously. I have argued so far that changes in material culture should be viewed through the lens of cultural evolution theory. This primarily means reconstructing the material records explicitly in terms of functional adaptations to culture-infused worlds, instead of e. g. supposedly non-functional body decorations. It further means favoring a gradualist approach at (nearly) all costs, no matter how rapid culture-infused environments may change under the right circumstances. Finally, it means viewing the material-archaeological records within a much larger history of sharing intentionality and transmitting culture, the bulk of which does not materialize.

7 A plea for stages

Donald’s (1991) strategy, as of many others, was to accept continuity, and to divide prehistorical cultural evolution into stages. Rather than asking the familiar “like us or not like us” questions of the old philosophical anthropology, thinking in stages means to provide “broad, unifying concepts that express the dominant cognitive quality of the individual mind in relation to society” (Donald 1991, 269). Such thinking in stages is compatible with a gradualist account of human cultural evolution which views products of material culture as functional adaptations to the culture-infused environments of the time.

Such cultural-evolutionary thinking should be part and parcel of the field of evolutionary cognitive archaeology. However, the evolution of culture is not a mere extension of genetic evolution. As I have argued, it operates on its own principles, and therefore the study of prehistoric cultural change demands a different kind of evolutionary theorizing. In particular, identifying the adaptive function of prehistoric items presupposes an understanding of the adaptive problems they are to solve. This is particularly complicated, because the environments which exerted selection pressure on prehistoric cultural items were themselves culturally infused, and largely in ways which do not materialize. For example, beadwork production may fix an adaptive problem that is itself already predominantly cultural, one that is mediated by complex social hierarchies and some symbolic speech, both of which depend on generative transmission. Beadwork might thus be an adaptation to a problem which is itself already cultural, and furthermore largely immaterial, as pre-existing hierarchies and communicative systems need not materialize. The generative transmission of beadwork would in turn “reprogram” future generations of hominins, thereby modifying the cultural worlds they inhabit, which then exert new cultural (as well as biological) pressures, and so on.

I have argued that, given that evolution is a gradual process, it makes no sense to fit all the data into a pre-modern/modern conceptual schema. Instead, more specific cultural-evolutionary hypotheses are needed, which can make gradual changes in material culture over longer periods of time understandable. The construction of such hypotheses will rely largely on reverse-engineering: taking the archaeological findings and other data as starting points, we need to reconstruct the problems these adaptations were selected to solve. It is not unusual in evolutionary biology to formulate (functional) hypotheses which may not be verifiable by current means. They can be scientific hypotheses nonetheless, and their credibility depends on how they explain or predict data in accordance with viable theoretical models.

8 Conclusion

I argued in this chapter that thinking about modern symbolism requires taking evolution seriously. I first outlined some of the problems involved in thinking about modern symbolism. These problems are substantial, which means archaeologists (and philosophers) should resist framing discoveries in terms of “modern symbolism”, as if something new is thereby learned about human behavioral complexity or cognitive evolution. I then defended a concept of culture primarily based on the psychological and phenomenological principles of shared intentionality, and outlined a theory of how cultures evolve. Finally, I returned to the question of symbolism, where I suggested that continuity thinking is to be preferred, and that stage-based approaches are the way forward, as they are compatible with gradualism, and avoid the old dichotomies. Rather than debating the “modern” nature of certain items, such stage-based thinking should focus on the construction of hypotheses concerning the adaptive problems which cultural items were selected to solve.